Screen-time influences children's mental imagery performance - Eliant.eu

Page created by Lance Barnes
 
CONTINUE READING
Screen-time influences children's mental imagery performance - Eliant.eu
Received: 7 August 2019    |   Revised: 12 March 2020   |   Accepted: 17 April 2020

DOI: 10.1111/desc.12978

PAPER

Screen-time influences children's mental imagery performance

Sebastian P. Suggate | Philipp Martzog

Department of Educational Sciences,
University of Regensburg, Regensburg,             Abstract
Germany                                           Mental imagery is a foundational human faculty that depends on active image con-
Correspondence                                    struction and sensorimotor experiences. However, children now spend a significant
Sebastian P. Suggate, Department of               proportion of their day engaged with screen-media, which (a) provide them with
Educational Science, Universitaetsstr. 31,
93040 Regensburg, Germany.                        ready-made mental images, and (b) constitute a sensory narrowing whereby input
Email: sebastian.suggate@ur.de                    is typically focused on the visual and auditory modalities. Accordingly, we test the
Funding information                               idea that screen-time influences the development of children's mental imagery with a
Software - AG Stiftung, Grant/Award               focus on mental image generation and inspection from the visual and haptic domains.
Number: ER-P 11657
                                                  In a longitudinal cross-lagged panel design, children (n = 266) aged between 3 and
                                                  9 years were tested at two points in time, 10 months apart. Measures of screen-time
                                                  and mental imagery were employed, alongside a host of control variables including
                                                  working memory, vocabulary, demographics, device ownership, and age of exposure
                                                  to screen-media. Findings indicate a statistically significant path from screen-time at
                                                  time 1 to mental imagery at time 2, above and beyond the influence of the control
                                                  variables. These unique findings are discussed in terms of the influence of screen-
                                                  time on mental imagery.

                                                  KEYWORDS

                                                  cognitive development, electronic media, mental imagery, mental simulation, screen-media,
                                                  screen-time

1 | I NTRO D U C TI O N                                                            hallmark features of screen-time—almost regardless of whether the
                                                                                   device is a television, smartphone, or computer—are, firstly, its degree
                                                                                   of passivity regarding its provision of mental images and, secondly,
        TV provides the viewer with ready-made visual im-                          its sensory narrowing. Beginning with passivity, the images provided
        ages and thus does not provide viewers with practice                       by screens can generally be described as “ready-made” in that they
        in generating their own visual images.                                     are provided directly via the screen media. Accordingly, it could be
                    (Valkenburg & van der Voort, 1994, p. 317)                     surmised that they may not require active image construction, other-
                                                                                   wise typical in mental life such as when reading text (i.e. the reduction
    Children spend a significant proportion of their time operat-                  hypothesis, see Valkenburg & van der Voort, 1994). Second, during
ing, viewing, and engaging with screen devices such as televisions,                screen-time sensory input is typically dominated by two modalities,
computers, game consoles, tablets and smartphones—sometimes                        namely the visual and auditory, presumably somewhat to the detri-
in excess of 4 hr/day (Gingold, Simon, & Schoendorf, 2014; Hinkley,                ment of others (e.g. tactile, proprioceptive, visceroceptive, and even
Salmon, Okely, Crawford, & Hesketh, 2012; Rideout, 2017). Two                      olfactive and gustative).

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in
any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2020 The Authors. Developmental Science published by John Wiley & Sons Ltd

Developmental Science. 2020;00:e12978.                                                                        wileyonlinelibrary.com/journal/desc   |   1 of 13
https://doi.org/10.1111/desc.12978
2 of 13   |                                                                                                                    SUGGATE and MARTZOG

    It is now indisputable that sensory experience provides human
cognition with not only input, but impetus for its development
                                                                                Research Highlights
(Lewkowicz, 2000), with mental imagery and thought depending
                                                                                • Mental imagery lies at the heart of mental life and re-
on activation of sensory areas extending beyond the visual and au-
                                                                                   quires both active image generation and a broad range
ditory modalities (Connell & Lynott, 2012; James, 2010; Martzog &
                                                                                   of sensorimotor experiences.
Suggate, 2019; Wellsby & Pexman, 2014). Although previous work has
                                                                                • Screen media provide children with ready-made and
considered passivity during screen-time (Valkenburg & van der Voort,
                                                                                   visually dominated mental images, hence may reduce
1994), research has neglected the question of whether screen-time's
                                                                                   multimodal mental imagery.
narrower sensory input affects an important aspect of cognitive de-
                                                                                • Using a longitudinal cross-lagged design with 266 chil-
velopment, namely, mental imagery. However, it is an open question as
                                                                                   dren we tested the effect of screen-time on mental im-
to whether screen-time suppresses mental imagery (i.e. reduction hy-
                                                                                   agery, controlling for a host of variables.
pothesis) or potentially stimulates imagery by training rapid processing
                                                                                • Greater screen-time linked to reduced mental imagery in
of images (i.e. the stimulation hypothesis). To address this tantalizing
                                                                                   children.
question, we use a longitudinal cross-lagged design to examine the ef-
fect of screen-time on children's mental imagery.

                                                                           the home environment, on features of cognitive development. The
1.1 | Active and passive screen-media                                      second research direction relates to work seeking to enhance chil-
                                                                           dren's learning and development by using media.
A number of studies have investigated links between screen-time and
aspects of cognitive, academic, and child development (Allchorne,
Cooper, & Simpson, 2017). Screen-time here is defined as time spent        1.2.1 | General cognitive development
viewing content displayed and projected from active and passive
screen-media, namely, those media that present visual information on       Despite attention-grabbing headlines such as “screentime is mak-
two dimensional displays. This encompasses traditional media (e.g. tel-    ing kids, moody, crazy and lazy” (Dunckley, 2015), research actually
evision and Personal Computers) and new media, such as smartphones         often lacks consistency of findings and often concrete theoreti-
and game-consoles. The defining feature of these media is that they        cal mechanisms linking screen-time to specific outcomes, particu-
convey sensory experiences primarily via the visual, but also auditory     larly in the case of mental imagery. Turning to findings, some have
senses, with only minor stimulation of other sensory modalities.           linked screen-time and eye-problems (Rosenfield, 2011), obesity
    Importantly for the current paper, the advent of active screen-me-     through impoverished physical movement (Fitzpatrick, Pagani, &
dia, such as touch-screens and media requiring direct interaction in       Barnett, 2012; Walsh et al., 2018), blue light exposure and sleep de-
shaping the course of subsequent media content (e.g. game-con-             ficiencies (Dworak, Schierl, Bruns, & Strüder, 2007), and academic
soles), requires careful consideration in comparison with more pas-        achievement through reduced time for formal learning (Beentjes
sive media (i.e. television viewing). First, these active screen-media     & van der Voort, 1988; Hancox, Milne, & Poulton, 2005; Weis &
generally require the input of participants in shaping the course of       Cerankosky, 2010).
the images provided by the screens. For example, the course taken             Turning specifically to cognitive development, findings are
in role-play computer games depends on active input, as does               mixed. Beginning with general developmental indicators, some re-
word-processing, taking photos, chatting, and so forth. Second, the        search indicates a small detrimental effect of excessive screen-time
haptic and fine-motor system is also active in delivering this input       on achieving developmental milestones (Madigan, Browne, Racine,
(e.g. pressing buttons), although in a comparatively impoverished          Mori, & Tough, 2019). Some studies also find that language develop-
form due to the sensory homogeneity of touch screens or keys (e.g.         ment in infancy is negatively affected by screen-time (Chonchaiya
Hipp et al., 2017). As discussed below, such active screen-media may       & Pruksananonda, 2008; Zimmerman, Christakis, & Meltzoff, 2007),
activate mental imagery in a different way to passive screen-media         others that young children do not acquire new words from screen
(i.e. television). For the purpose of the current paper, we exclude        media (Krcmar, Grela, & Lin, 2007; Robb, Richert, & Wartella, 2009),
new media, such as 3D interactive technologies, because these are          while others still demonstrate that educational content can impart
not in widespread use and little research exists on these.                 vocabulary (Rice, Huston, Truglio, & Writhgt, 1990) and narrative
                                                                           skill (Linebarger & Piotrowski, 2009; Linebarger & Vaala, 2010;
                                                                           Linebarger, 2005). However, an understanding as to why screen-
1.2 | Screen-time and cognitive development                                time might differentially affect language development is incomplete,
                                                                           although studies suggest that transferring from virtual to real worlds
At a general level, research on the effect of screen-time on cognitive     can be difficult for infants (Hipp et al., 2017).
development includes two sets of studies. The first group concerns            Executive functions have been examined more extensively.
broad effects of screen-time, usually for entertainment purposes in        One set of findings suggests that screen-time, particularly in the
SUGGATE and MARTZOG                                                                                                                      |   3 of 13

form of interactive video games, can enhance cognitive control in          argue in the next section, research has perhaps overlooked one key
adults (Anguera et al., 2013; Powers, Brooks, Aldrich, Palladino, &        feature of screen-media: such media present children with rapidly
Alfieri, 2013). Other studies find that the multitasking and rapid         changing pre-made sensory images that are typically specific to the
changes inherent in screen-time negatively affect executive func-          visual and auditory modalities alone. This might in turn influence the
tions in both adults (Ophir, Nass, & Wagner, 2009) and children            mental simulation of external events (i.e. mental imagery).
(Lillard & Peterson, 2011; Nathanson, Aladé, Sharp, Rasmussen,
& Christy, 2014). Furthermore, screen-time has been linked to in-
creased symptomology associated with attention-deficit hyperac-            1.3 | Mental imagery
tivity disorder (Nikkelen, Valkenburg, Huizinga, & Bushman, 2014).
However, further confusing the picture, a cross-sectional study            Visual imagery has been described as seeing with the mind's eye
from China found that screen-time positively linked to preschool           (Kosslyn, 1994) and the close cousin thereof, namely mental imagery,
children's executive functions (Yang, Chen, Wang, & Zhu, 2017).            can be understood as simulation or internal re-creation of percep-
Another study using a large Australian sample found that media             tual experience (Barsalou, 1999). Mental imagery can be conceived
exposure at age 2 years, but not age 4, negatively related to later        of as containing four stages, image generation, maintenance, scan-
self-regulation (Cliff, Howard, Radesky, McNeill, & Vella, 2018).          ning, and transformation (Kosslyn, Margolis, Barrett, Goldknopf, &
                                                                           Daly, 1990). Developmental effects also exist, with children being
                                                                           relatively poorer at generating, scanning, manipulating, or transform-
1.2.2 | Enhancing learning through screen-time                             ing images (Kosslyn et al., 1990). In addition, studies have shown that
                                                                           sensory and motor systems underlie mental imagery (e.g., Martzog &
On the other hand, a raft of approaches and studies demonstrate            Suggate, 2019), as has been found in other domains such as memory,
that, depending on age and content, children and adults can success-       language, and thought (Barsalou, 2008).
fully learn from screen-media (Barr & Linebarger, 2017). Generally,           Indeed, various theoretical approaches argue that sensory and
these approaches seek to capitalize on and enhance learning pro-           sensorimotor experiences form the basis of mental imagery and cog-
cesses via a number of techniques, sources, and strategies (Troseth,       nition. For instance, in embodied cognition theories, cognitive pro-
Strouse, Flores, Stuckelman, & Russo Johnson, 2020). When the              cesses have been described as resulting from an internal simulation
goals are clear and the program is well-designed, even passive             of underlying actions and perceptions (Barsalou, 2008; Glenberg
media (i.e. television) can enhance school readiness, problem solv-        & Gallese, 2012; Glenberg et al., 2008). According to perceptual
ing, and learning (see Kirkorian, Wartella, & Anderson, 2008). At a        symbols theory, Barsalou (1999) characterizes simulations as “the
theoretical level, well-designed programs could invoke established         top-down activation of sensory-motor areas” (p. 641). Re-enacted
learning principles, such as social learning, capturing and sustain-       perceptual experiences appear to bear close similarities to the ex-
ing attention, encouraging mental model development, reinforcing,          periences behind mental imagery (Kosslyn, 1994). Both simulation
facilitating explorative learning and knowledge elaboration (Barr &        theory (Jeannerod, 2001), and emulation theory of representation
Linebarger, 2017; Hattie, 2012; Kirkorian et al., 2008).                   (Grush, 2004), make the claim that motor and visual images are anal-
   Specifically, prompts provided by interactive electronic books          ogous to real-world physical and visual experiences because they
can support learning (Strouse & Ganea, 2016, 2017), especially for         make use of similar neural infrastructure as indicated by motor and
low SES families (Troseth et al., 2020) and including a social model       visual cortex activation during imagery (see Jeannerod, 2001 for a
(e.g. a parent) as a co-viewer can further enhance gains (Strouse,         review and Kosslyn, Ganis, & Thompson, 2001; Tomasino, Werner,
Troseth, O'Doherty, & Saylor, 2018). Pertinent for the current line        Weiss, & Fink, 2007).
of inquiry, children can experience difficulty deriving three-dimen-
sional information from two-dimensional media, which implicates
both children's sensorimotor systems and, as discussed next, mental        1.4 | Links between screen-time and
imagery in learning and cognitive development (Troseth, 2010).             mental imagery

                                                                           Although not directly investigating links between screen-time
1.2.3 | Summary                                                            and mental imagery, as here defined, there have been studies on
                                                                           links between television and day-dreaming/creative imagination
Indeed, taking research on screen-time and cognitive development           (Valkenburg & Peter, 2013; Valkenburg & van der Voort, 1994, 1995).
as a whole, we, along with others (e.g., Troseth, 2010), note that clear   Consistent with the reduction hypothesis, studies have found that
and plausible theoretical mechanisms need to be carefully tested           children perform more poorly on measures of creative and divergent
with ecologically valid designs amenable to causal interpretation,         production after viewing a television versus hearing a radio pro-
namely longitudinal cross-lagged panel designs (Kearney, 2017).            gram (Valkenburg & Beentjes, 1997). Furthermore, in a study with a
Effects appear to be context (Hirsh-Pasek et al., 2015) and develop-       large sample of children and using a cross-lagged design, television
mentally dependent (e.g., Barr & Linebarger, 2017). However, as we         viewing affected both the content of day-dreaming and reduced its
4 of 13   |                                                                                                                 SUGGATE and MARTZOG

occurrence (Valkenburg & van der Voort, 1995). On the other hand,             time at time 1, and screen time at time 2 from mental imagery at time
consistent with the stimulation hypothesis, rapid processing of               1, while accounting for control variables. A pattern consistent with
screen-images might stimulate the perceptual system (Dye, Green,              the unidirectional causal operation of screen-time on mental imag-
& Bavelier, 2009), perhaps explaining why some work has found indi-           ery would be found if the diagonal pathway from screen-time (t1) to
cations that video-gaming can support information processing (Dye             mental imagery (t2) were significant, but the converse pathway were
et al., 2009; Powers et al., 2013).                                           not, indicating unidirectional influences as opposed to bidirectional
                                                                              association (Kearney, 2017).
                                                                                 Furthermore, the design permitted us to control for a host of the-
1.5 | The current study                                                       oretically important control variables beyond parental demographic
                                                                              data and including working memory and vocabulary. The latter two
As outlined and defined here, two features define mental im-                  are key covariates because working memory is intimately related to
agery. First, mental imagery constitutes activity in the form of              executive functions, which, along with vocabulary, have been found
image generation, maintenance, scanning, and transformation                   to relate to screen-time usage. Additionally, we used a novel men-
(Kosslyn et al., 1990). Second, mental imagery depends on broader             tal imagery measure, namely a mental comparison task, designed to
sensorimotor simulations and experiences (Barsalou, 1999;                     specifically target the sensorimotor foundations of mental imagery
Kosslyn, 1994). Two functional properties of screen media bear                (Martzog & Suggate, 2019), that generates response accuracy and
close examination.                                                            response latency.
    First, screen-media provide images that presumably somewhat                  In accordance with the reduction hypothesis, we expected that
circumvent the effortful construction processes required during               mental imagery performance (accuracy) would be lower in children
mental imagery, which has been called the reduction hypothesis                exposed to more screen-time because they have less experience
(Valkenburg & van der Voort, 1994). Conceivably, various screen-me-           with the active creation of their own mental images. Theoretically, it
dia might differentially result in a reduction in mental imagery, for         is conceivable that screen-media train rapid processing of mental im-
instance if during screen-time children anticipate, or reflect on, con-       ages that have been provided by screens, perhaps leading to greater
tent, then some mental imagery might be employed. Furthermore, if             mental imagery processing speeds for familiar images. Additionally,
actions are to be planned and executed via screen-media, it is likely         previous work has found that screen-time increases children's impul-
that mental imagery of subsequent actions would be stimulated                 sivity (Lillard & Peterson, 2011; Nikkelen et al., 2014) and process-
more so with active than passive media. Accordingly, it might be ex-          ing speed (Dye et al., 2009). Accordingly, we tentatively expected
pected that screen-media generally would reduce the active gener-             screen-time to result in faster response latencies, consistent with
ation of mental images, but that active screen-media might have less          the stimulation hypothesis (Valkenburg & van der Voort, 1994).
of a detrimental effect on imagery compared to passive media.                 Finally, we tested whether passive and active screen-media differ-
    Second, as mentioned, screen-time represents a sensory narrow-            entially relate to mental imagery, reasoning that the added activity
ing, in that two modalities (i.e. the visual and auditory) are likely stim-   (i.e. planning and executing actions) inherent in active screen-media
ulated while other broader sensorimotor experiences (e.g. motor,              means that active screen-time may not relate negatively to mental
haptic, proprioceptive) are suppressed. Again, it may be important            imagery.
to distinguish between passive and active screen-media, in that the
latter involve some direct motor action (Galetzka, 2017). However,
given the homogeneity of motor action when interacting with flat              2 | M E TH O D
screens or analogous buttons—which by nature vary little in terms of
haptic or proprioceptive feedback—screen time likely results in com-          2.1 | Participants
paratively impoverished sensorimotor experiences otherwise to be
expected in childhood (e.g. outdoor play, block games). Indeed, men-          Participants in this study were 266 children (51.1% girls) aged be-
tal imagery, which is itself a fundamental building block of thought          tween 35 and 120 months (M = 75.26, SD = 20.05) at the first time
and cognition (Barsalou, 1999; Kosslyn, Ganis, & Thompson, 2003),             point, attending either preschools (n = 141) or primary schools
depends on both broader sensorimotor experience and active image              (n = 125) in a small city in Germany. Thirty-two percent of children
generation. Thus, it would appear pressing to investigate the effect          had at least one parent born in a foreign country and 26.3% spoke
that children's screen-time experiences have on their mental imag-            a language other than German at home. Aside from German, there
ery performance. However, to date, no study has directly investi-             was no clear majority amongst the home languages spoken, with
gated this.                                                                   these being a mixture of Eastern European and Asiatic languages.
    Accordingly, we conducted a longitudinal cross-lagged panel               Additionally, 56.4% of the families had at least one parent with a
study measuring 266 preschool and primary school children's mental            University degree or equivalent. The national average for individual
imagery and screen-time use at two points in time, 10 months apart.           adults (and hence not directly comparable) in a similar age range to
Crucially, our use of a cross-lagged panel design has the key advan-          the parents is 29% for this generation (Federal Bureau of Statistics,
tage that mental imagery at time 2 can be predicted from screen               2017), indicating that the sample was likely more highly educated.
SUGGATE and MARTZOG                                                                                                                      |   5 of 13

2.2 | Measures                                                           smartphone, game-console), giving a theoretically possible score
                                                                         range from 4 to 32.
Demographic data were collected via a parent questionnaire.
Parents were asked about languages spoken at home, educational
background, home country, screen-time usage, device ownership,           2.2.2 | Mental imagery
and first contact with media.
                                                                         We employed a mental imagery task based on previously used men-
                                                                         tal size comparisons tasks (Moyer, 1973; Paivio, 1975), that has been
2.2.1 | Screen-media questionnaire                                       recently utilized and further developed (Martzog & Suggate, 2019).
                                                                         Pertinent to the task was that children needed to rely on information
A parent questionnaire was used to measure children's screen-            derived from the mental images themselves, not declarative knowl-
time and media usage. Given notorious difficulties in measuring          edge about the images. Children were asked to imagine two specific
screen-time, in part due to information bias and social desirability,    objects, and then asked to make a judgment as to which from the tar-
we optimized our method over the course of several pilot stud-           get and distractor item was better encapsulated by a sensory feature
ies. At a theoretical level, measures involving diaries have been        (i.e. “which is shinier, [a] trumpet or [a] violin?”). Note that the ques-
recommended in preference to questionnaires because these are            tion was thus phrased, such that the stimuli appear at the end so that
thought to provide more accurate estimates (Reinsch, Ennemoser,          processing can only begin after presentation. Also, in German, the
& Schneider, 1999). However, one key disadvantage with diaries           indefinite article “a” was not grammatically necessary in the ques-
is low-compliance. To address these issues, we opted for a diary-        tion sentence, thus reducing memory load between presentation of
questionnaire format in which parents were asked about their chil-       the two target stimuli. The invoked modalities were determined by
dren's screen time activities at different points in the day. Thus,      two conditions, firstly the question contained an adjective pertain-
during the working week, we asked about usage before school/             ing to the modality (e.g. “shiny”) and secondly, the target words had
preschool, in the afternoon, and in the evening, and then on the         a sensory feature that was prominent in the corresponding modality
weekend. Additionally, we asked about the amount of time spent           (e.g. “trumpet”). Although the original task contained stimuli pertain-
on various devices, including televisions, computers, tablets, play-     ing to the haptic, visual, and visual-haptic modalities, analyses found
consoles, and smartphones. Thereby, we responded to previous             that the task was best conceptualized as a general imagery measure
work calling for a focus on more modern media in addition to tele-       (Martzog & Suggate, 2019). During task development, Martzog and
vision (Valkenburg & Peter, 2013). Because of our hypotheses that        Suggate (2019) accounted for diverse lexical features (e.g. length,
screen-time affects imagery via sensory-narrowing, we also asked         syllabic structure, frequency, imageability, manipulability, sensory
parents how old children were when they first began using the            ratings).
various appliances to determine the effect of long-term exposure.            Response accuracy and latency were both recorded by the ex-
Finally, we also included questions asking about the ownership of        perimenter using response keys on a laptop. In total there were 34
electronic media appliances.                                             items, each containing a distractor and a target and children were
   Accordingly, our data provided three scores: (a) device owner-        asked to respond as soon as they knew the answer without, how-
ship, (b) daily exposure time, and (c) age at which exposure began.      ever, emphasizing speed in order to avoid hectic and erratic re-
Screen-time was rated on a 6-point Likert scale for each medium (no      sponses. Response accuracy and response time was recorded and
screen-time,
6 of 13   |                                                                                                               SUGGATE and MARTZOG

2.2.4 | Vocabulary                                                        TA B L E 1 Descriptive statistics for control variables, screen-
                                                                          time, and mental imagery

Children's vocabulary was assessed using the vocabulary test at time                             Descriptive statistics
1 from the Kaufmann ABC (Kaufman & Kaufman, 2015). In this task,
                                                                           Variables             M        SD         N       Min     Max
children are shown pictures and are required to name the object in
the pictures. One point was awarded for each correct item and there        Control variables

was a discontinue rule after 4 consecutive errors, and a basal item          Age (months)        75.26    20.05      259     35      120
was established after three correct responses. The maximum num-              Vocabulary          19.02    5.28       255      4.00   37.00
ber of points possible was 39 and the internal consistency of the            Working             2.44     1.28       248      0      4.00
vocabulary test was estimated at α = 0.89.                                    memory
                                                                             Screen exposure     23.18    3.90       255     11.00   32.00
                                                                              (age)

2.3 | Procedure                                                              Device              5.68     1.173      255      1.00   7.00
                                                                              ownership

Children were tested twice on the screen-time, imagery, and some           Time 1 variables

of the control variables, on average 9.81 (SD = 1.33) months apart,          Mental imagery      26.86    5.36       254      2.00   34.00
                                                                              (acc)
once in the academic year of 2017–2018 and again in 2018–2019, in
                                                                             Mental imagery      2,589    1,068      254     745     10,023
their educational institutions. Data were drawn from a larger longi-
                                                                              speed (ms)
tudinal study in progress. All tasks were administered individually by
                                                                             Screen-time         1.87     1.43       237      0.00   9.14
trained research assistants and the second author according to test
                                                                           Time 2 variables
instructions. Parents completed questionnaires, at two time points
parallel to data collection, providing information on their children's       Screen-time         1.52     1.11       197      0.00    5.93

screen-time and demographic data. For preschool children, between            Mental imagery      25.62    4.97       250      2.00    32.00
                                                                              (acc)
two and three testing sessions were required, each of approximately
                                                                             Mental imagery      2,327    1,494      250     487      13,105
20 min, so as to not overtax concentration spans. Approval from the
                                                                              speed (ms)
Ministry of Education was obtained prior to conducting the study
and written consent was provided by the parents of participating
children, followed by the latter's verbal assent.                            Finally, although in our mental imagery task we did not directly
                                                                          ask children to respond as quickly and accurately as possible, the
                                                                          data represent a double challenge in that response time is not inde-
2.4 | Data analyses                                                       pendent of response accuracy. Accordingly, we treated these two
                                                                          variables separately, reasoning that response speed—regardless of
Data analyses consisted of first screening the data for anomalies         accuracy—provided one source of information about how partici-
(skew and kurtosis) and calculating descriptive statistics. We win-       pants approached the task (i.e. the stimulation hypothesis) and that
sorized the data by capping outliers on mental imagery reaction time      response accuracy to another (i.e. the reduction hypothesis).
to the three standard deviations above the item level mean. Next,
we conducted correlation analyses to explore relations between the
exogenous and endogenous variables central to the cross-lagged            3 | R E S U LT S
panel design and path modeling. Path models allowed us to conduct
regression analysis (Kline, 2011) testing for cross-lagged effects con-   3.1 | Descriptive statistics
sistent with the causal operation of screen-time on mental imagery
(Reinders, 2006), controlling for the influences of a host of varia-      In Table 1 the descriptive statistics for scores on the screen-time,
bles (Byrne, 2010; Kline, 2011). Path models were calculated using        mental imagery, and control variables are presented. Inspection of
AMOS v. 23 (Arbuckle, 2014) with missing values being given full          skew and kurtosis statistics suggested that data were normally dis-
consideration through full maximum likelihood estimation. Screen-         tributed, however, response latency to the imagery items appeared
time and mental imagery were modeled as manifest variables to fa-         to be right skewed (skewedness in the vicinity of 2.50). Of the im-
cilitate model convergence and the control variables (presented in        agery data, 4.5% was missing at time 1 and 6.0% at time 2, with the
Table 2) were added as predictors with paths onto both time 1 and         corresponding percentages for the screen-time data being 10.9%
time 2 screen-time and mental imagery. Control variables included         and 25.94%. Next correlation coefficients were calculated for the
parent education, ethnic status, device ownership, age of exposure        variables in Table 1, which are presented in Table 2. Trends indicate
to screen-media, vocabulary, working memory, chronological age,           that screen-time correlated negatively with accuracy, as did re-
and, to control for variations in testing procedure, the number of        sponse speed on the imagery task. Vocabulary and working memory
months between time 1 and time 2.                                         positively predicted mental imagery and were generally negatively
SUGGATE and MARTZOG                                                                                                                                  |    7 of 13

TA B L E 2       Product-moment and partial correlation coefficients between screen-time, control, and mental imagery variables

                                1           2           3            4           5            6           7           8             9                10

 1          Vocabulary               —          0.38*       0.08         0.28*   −0.12            0.55*   −0.25*      −0.21*            0.36*        −0.19*
 2          Working memory          0.61*        —          0.08         0.19*   −0.14*           0.34*   −0.15*      −0.02             0.23*        −0.02
 3          Screen exposure         0.05        0.04         —       −0.13*      −0.36*           0.13*       0.01    −0.27*            0.14*        −0.05
             (age)
 4          Device ownership        0.30*       0.22*   −0.13*            —          0.06         0.17*       0.01    −0.06             0.22*        −0.04
 5          Screen-time t1          0.02        0.05    −0.35*           0.08         —       −0.05           0.01        0.64*     −0.21*           −0.02
 6          Mental imagery          0.71*       0.63*       0.08         0.20*       0.10          —      −0.26*      −0.12             0.35*        −0.20*
             t1 (acc)
 7          Mental imagery      −0.34*      −0.29*          0.01     −0.02       −0.04        −0.36*           —      −0.01         −0.12                0.27*
             speed t1 (ms)
 8          Screen-time t2      −0.04           0.15*   −0.27*       −0.03           0.66*        0.07    −0.07            —        −0.17*           −0.04
 9          Mental imagery          0.54*       0.51*       0.10         0.25*   −0.06            0.58*   −0.24*      −0.01              —           −0.03
             t2 (acc)
 10         Mental imagery      −0.34*      −0.26*      −0.04        −0.08       −0.09        −0.37*          0.33*   −0.13         −0.21*                —
             speed t2 (ms)
 11         Age t1                  0.55*       0.69*   −0.03            0.12        0.21*        0.65*   −0.26*          0.24*         0.53*        −0.35*

Note: Correlations above the diagonal have age partialled out, t1 = time 1, t2 = time 2.
*p < .05.

     Screen-
                                            res.                                                                                             res.
     media
     (capital)
     Age-                             Mental imagery                                                                              Mental imagery
     exposure                              (t1)                                       β = .22*                                         (t2)
     screen-time
     Chronologi
     cal age
     Ethnic
                                                                                             β = -.16*
     status
     Time
     between t1
     and t2                                                                                        β = .00
     Parental
     education
     Vocabulary                       Screen-time (t1)                                 β = .59*                                   Screen-time (t2)
     Working
     memory

                                            res.                                                                                              res.

FIGURE 1          Structural equation model depicting cross-lagged panel design testing links between screen-time and mental imagery
8 of 13     |                                                                                                                  SUGGATE and MARTZOG

TA B L E 3      Estimates for influence of control variables on screen-time and mental imagery from structural equation model in Figure 1

                           Screen-time t1               Mental imagery t1                    Screen time t2                Mental imagery t2

 Variable                  B       SE       β           B       SE          β                B        SE      β            B         SE        Β

 Age (months)               0.02   0.01         0.24*    0.09   0.02            0.34*         0.01    0.01        0.14       0.06    0.02      0.23*
 Vocabulary                −0.01   0.02     −0.08        0.40   0.06            0.39*        −0.04    0.02    −0.20*         0.15    0.07      0.17*
 Working memory             0.00   0.10     −0.00        0.64   0.26            0.15*         0.11    0.08        0.12       0.36    0.30      0.09
 Screen exposure           −0.11   0.02     −0.31*       0.07   0.06            0.05         −0.03    0.02    −0.11          0.00    0.07      0.00
  (age)
 Device ownership           0.06   0.07         0.05     0.17   0.20            0.04         −0.03    0.05    −0.03          0.48    0.22      0.11*
 Ethnic status              0.22   0.19         0.07    −0.45   0.50        −0.04             0.15    0.14        0.06       0.68    0.58      0.07
 Parental                  −0.43   0.18     −0.15*       0.54   0.46            0.05         −0.01    0.13    −0.01        −0.01     0.53      0.00
  education

*p < .05.

associated with screen-time. Age of media exposure was correlated               as predictors in the model, and covarying the imagery residuals. The
with screen-time and device ownership.                                          model again showed good global fit, χ2/df = 1.54, CFI = 1.00, and
                                                                                RMSEA = 0.05, but the path of interest between screen-time and
                                                                                mental imagery response latency—although in the direction pre-
3.2 | Cross-lagged effect between screen-time and                               dicted by the stimulation hypothesis—was not statistically signifi-
mental imagery                                                                  cant, β = −0.05, p = .46, nor was the converse path from imagery to
                                                                                screen-time, β = −0.01, p = .91. In an alternative procedure, we (nat-
We estimated two models to test links between screen-time and                   ural) log transformed the response latencies, which transformed the
mental imagery, one each for mental imagery accuracy and mental                 skew and kurtosis statistics to near zero for these data; however, the
imagery speed. In both instances, the control variables were speci-             cross-lagged path from screen-time at time 1 to imagery response
fied to predict the screen-time and mental imagery variables, which             latency at time 2 was not significant, β = −0.06, p = .37, despite good
contained cross-lagged paths. Beginning with accuracy, the model                model fit.
converged on the 9th iteration. A chi-square value estimating good-                Finally, we examined the possibility of a speed accuracy trade-off
ness of fit was not significant, χ2(2) = 1.34, p = .51, indicating good         operating, whereby participants' response latencies were shorter at
model fit. In addition to the chi-square statistic, other fit indices           the expense of greater accuracy (Heitz, 2014). Product moment cor-
are recommended, namely, that the CFI should be around or above                 relation coefficients indicated that accuracy correlated negatively
CFI = 0.95, RMSEA around or below 0.05, and that χ2/df should not be            with speed, r = −0.40 at time 1 and r = −0.21 at time 2, ps < .002,
significant against a chi-square distribution (Byrne, 2010). The cur-           thus suggesting the opposite of a speed-accuracy trade-off because
                                                   2
rent estimates indicated excellent model fit, χ /df = 0.67, CFI = 1.00,         faster responders were more accurate. To control for developmen-
and RMSEA = 0.000 (Byrne, 2010; Kline, 2011). As can be seen in                 tal influences, the partial correlation coefficients controlling for age
Figure 1, accounting for the host of control variables screen-time              between response latency and accuracy were calculated. At time 1,
at time 1 negatively predicted mental imagery at time 2, whereas                this was negative and significant, r(248) = −0.28, p < .001), indicat-
the converse was not true. In Table 3, working memory, vocabulary,              ing that greater accuracy was linked to greater speed, however, this
and chronological age were significant predictors of mental imagery.            correlation was not significant at time 2, r(240) = 0.00, p = .95. Thus,
Both chronological age and age of exposure to screen media also                 although the data do not indicate a speed-accuracy trade-off, the
predicted screen-time. To estimate the correlation between screen-              previous model was re-run, this time including accuracy as a con-
time and mental imagery accuracy at time 1, their residuals were                trol variable, however, this did not alter the magnitude of the small,
covaried. The corresponding correlation was not statistically signifi-          negative, but nonsignificant path between screen-time and mental
cant, r = 0.08, p = .23.                                                        imagery response latency.
    In a second path model, the same model was used with the ex-
ception that response latency replaced response accuracy in the im-
agery task. The model again showed good global fit, χ2/df = 0.64,               3.3 | Passive versus active screen-media and
CFI = 1.00, and RMSEA = 0.000, but the path of interest, between                mental imagery
screen-time at time 1 and mental imagery response latency at time
2, was not statistically significant, β = −0.06, p = .35, nor was the           To investigate links between active versus passive screen-media and
converse path from imagery to screen-time, β = −0.01, p = .93.                  mental imagery we first examined descriptive statistics pertaining to
Additionally, we attempted to partial out the influence of accuracy             the daily engagement with the various media. These are presented
from the mental imagery reaction time measurements by using these               in Table 4. As can be seen in Table 4, television constituted the most
SUGGATE and MARTZOG                                                                                                                      |   9 of 13

heavily used screen-medium in this sample. The remaining media            pathways without sacrificing ecological validity as in laboratory ex-
were scarcely used and, given that they can all be classified as ac-      periments (Kline, 2011).
tive screen-media, were aggregated into a single measure (i.e. ac-           Findings were clear in three regards. First, children who spent
tive screen-media) for subsequent analysis. Next two structural path      greater amounts of time using screen media showed statistically
models were calculated, replicating those presented in Figure 1,          significantly lower performances on mental imagery accuracy,
with the exception that one was calculated for active and the other       consistent with the reduction hypothesis (Valkenburg & van der
for passive screen-time. As shown in Table 5, both models fitted          Voort, 1994). Thus, our hypothesis that viewing screen-media, by
the data well and, although the models are not depicted in full due       virtue of their providing ready-made mental images that suppress
to space constraints, they were highly similar to those in Figure 1.      active image generation, receives initial support.
Importantly, both types of screen-media at time 1 showed a similar,          Second, we found virtually no difference in the negative cross
statistically significant, cross-lagged link to mental imagery accuracy   lagged-link between screen-time and mental imagery for media
at time 2.                                                                classified as active versus passive. On the one hand, this finding is
                                                                          surprising because we expected that more active media would in-
                                                                          volve mental imagery abilities to a greater extent. However, our sup-
4 | D I S CU S S I O N                                                    position has not been verified by empirical evidence, such that it is
                                                                          plausible that even many more so-called active media types might
We tested, for the first time, whether children's mental imagery          still not involve much active imagery generation, perhaps especially
abilities were affected by screen-time, the latter of which now           in comparison to other typical childhood experiences (e.g. reading,
constitutes a significant proportion of the mental activities that        imaginative play).
children engage in Gingold et al. (2014), Hinkley et al. (2012) and          Third, contrary with the stimulation hypothesis, screen-time did
Rideout (2017). Two features of screen-time that have scarcely been       not relate to children's response latencies on the mental imagery
investigated are, first, its sensory narrowing (i.e. dominance of the     task, as we had posited based on previous work (Dye et al., 2009;
auditory-visual modalities) and, second, its providing ready-made         Lillard & Peterson, 2011; Nikkelen et al., 2014). Perhaps the dosage
and often rapidly changing images which potentially suppress the          of screen media here, which was less than in previous work, may
active mental life (Valkenburg & van der Voort, 1994). We reasoned        not have been sufficient to induce the greater impulsivity neces-
that these two features of screen-time might lead to negative asso-       sary to affect response latencies. In terms of our suggestion that
ciations with mental imagery accuracy via the reduction hypothesis        screen-media might train the perceptual system (Dye et al., 2009),
and, conversely, a decrease in response latencies as predicted by the     with hindsight, it could instead be argued that this likely only applies
stimulation hypothesis. Furthermore, we reasoned that different (i.e.     for certain kinds of games unlikely to have been systematically em-
active vs. passive) screen-media might differentially affect mental       ployed here—especially given that the dominant form of screen-time
imagery. Finally, we tested these hypotheses using a longitudinal         in this sample was television viewing. Accordingly, we tentatively
cross-lagged panel design, which has the advantage of testing causal      conclude that screen-time does not stimulate mental imagery per-
                                                                          formance when this requires mental comparisons of visual/haptic
TA B L E 4   Estimated daily (hours) time spent with various media
                                                                          images.
devices
                                                                             Alongside screen-time, vocabulary, working memory, and chrono-
                       Descriptive statistics                             logical age were also significant predictors of mental imagery. At a con-
                                                                          ceptual level, our mental comparisons task required working memory
 Variables             M        SD       N          Min         Max
                                                                          because the participants were required to compare mental images
 Time 1
                                                                          to solve the task. Accordingly, we controlled in advance for work-
    Television         1.01     0.84     237        0.00        4.71
                                                                          ing memory. Although it might be tempting to apply a similar line of
    Active media       0.17     0.20     237        0.00        1.14      reasoning to vocabulary's influence on mental imagery, we consider
    PC                 0.06     0.20     237        0.00        1.64      it unlikely that children's vocabulary knowledge directly constrained
    Smartphone         0.18     0.45     237        0.00        4.57      task performance. Specifically, although it is true that children would
    Tablet             0.33     0.50     237        0.00        3.07      have to know the words in the mental comparisons task in order to
    Gaming console     0.10     0.32     237        0.00        2.14      be able to image them, stimuli were simple and hence could be ex-
 Time 2                                                                   pected to be familiar to the children (see Martzog & Suggate, 2019 for

    Television         0.76     0.67     197        0.00        3.93      stimuli and a discussion of this task). Instead, we suggest that children
                                                                          with larger vocabularies likely have richer perceptual representations
    Active media       0.20     0.25     195        0.00        1.30
                                                                          in general (Connell & Lynott, 2016; Hargreaves, Pexman, Johnson, &
    PC                 0.09     0.29     192        0.00        2.14
                                                                          Zdrazilova, 2012; Suggate & Stoeger, 2017), which leads to greater
    Smartphone         0.21     0.41     192        0.00        2.36
                                                                          imagery performance. Support for this idea also comes from the con-
    Tablet             0.32     0.48     194        0.00        3.29
                                                                          tribution of age to mental imagery performance found here, which
    Gaming console     0.16     0.39     193        0.00        2.29
                                                                          suggests that a more mature cognitive system relates to performance,
10 of 13    |                                                                                                              SUGGATE and MARTZOG

TA B L E 5      Model parameters comparing cross-lagged path for active versus passive screen-time (time 1) on mental imagery (time 2)

                                                                                                                              Screen-time to
 Model                          df            X2             p              χ 2/df        CFI                RMSEA            mental imagery (β)

 Active screen-media            2             0.10           0.95           0.05          1.00               0.00             −0.12*
 Passive screen-media           2             2.66           0.26           1.33          0.99               0.04             −0.11*

*p < .05.

extending beyond specific lexical level knowledge directly derived        4.2 | Limitations and future work
from the imagery items.
                                                                          In the current study, we found that children spent, on aver-
                                                                          age, nearly 2 hr/day engaged in screen-media usage. This figure
4.1 | Theoretical and practical implications                              is consistent with previous work for this age group in Germany
                                                                          (Feierabend, Plankenhorn, & Rathgeb, 2017), but is still somewhat
The current study adds to the rapidly growing body of research            lower than that found in the United States, for example (Gingold
looking at children's learning and development in relation to screen-     et al., 2014), although more recent data from the United States also
media (e.g., Herodotou, 2018). In terms of developmental work, the        found a mean daily screen-media usage of 2 hr and 19 min. Reasons
study's findings contribute to work suggesting that screen-time af-       for this difference are speculative, but might be due to the region
fects child development in a complex manner, with mental imagery          in which the study was conducted, which has rural surroundings,
now seemingly a factor to consider amongst others (see Barr &             low levels of crime, and a culture in which unsupervised outdoor
Linebarger, 2017).                                                        play is still encouraged. Presumably, conducting the study in sam-
    According to the current data, the, on average, 1 hr of televi-       ples with greater levels of screen-time would result in greater as-
sion viewing per day (ranging up to a maximum of 4 hr and 42 min)         sociations with mental imagery due to reduced opportunities to
across the course of nearly 10 months was enough to negatively            engage in compensatory activities for the effects of screen-time.
affect mental imagery accuracy at time 2. More surprisingly, the          In a similar vein, in terms of statistical variance, such work might be
corresponding time 1 engagement with active screen-media of               especially fruitful in the United States where children spend about
up to a maximum of 68 min/day—with a sample average of just               twice as much time interacting with smart-phones (Rideout, 2017).
10 min/day—was enough to negatively predict time 2 mental im-                  Children's daily experiences appear to increasingly include
agery. Turning these figures into total exposure across the course        screen-time experiences, which may come at the expense of time
of the study, across 10 months, it is likely that children spent, on      engaged in activities that require greater levels of mental imagery
average, over 300 hr watching television—with the heavy viewers           (Ennemoser & Schneider, 2007; Weis & Cerankosky, 2010), such
spending around 1,410 hr. In terms of the reduction hypothesis,           as reading (Glenberg, Brown, & Levin, 2007) or imaginative play
perhaps then it is not surprising that links between screen-time          (Wallace & Russ, 2015). Accordingly, future work could expand on
and mental imagery were found. Accordingly, mental imagery                the current findings and test whether home reading activities, for
seems to undergo continual development in the age of samples              example, offset effects of screen-time and support mental imagery
studied here and seems sensitive to reduced practice at the active        development. In a similar vein, future work needs to examine the
generation of mental images. As such, the current work is consis-         sensorimotor consequences of screen-time. In the current paper, we
tent with studies showing that mental processes and concepts are          allude to a sensory narrowing during screen-time, in that the visual
dependent on a rich array of sensorimotor information and pro-            and auditory senses are stimulated whereas other sensorimotor mo-
cesses (Connell & Lynott, 2016; Hargreaves et al., 2012; Suggate          dalities may be neglected (i.e. proprioception, active motor control,
& Stoeger, 2017).                                                         olfaction, gustation, haptics). Accordingly, future work should test
    The current study also extends previous work on media learn-          sensorimotor development as a function of screen-time and time
ing. Research has examined the conditions under which screen-me-          engaged in sensorimotor activities, also studying neuroanatomical
dia contribute to learning, among other factors, focusing on media        changes underlying these skills (see Hutton, Dudley, Horowitz-
content presentation, and children's developmental readiness              Kraus, DeWitt, & Holland, 2019).
(e.g., Barr & Linebarger, 2017). Although some work has examined               In the current study, although we examined a host of control vari-
the medium itself, for example by comparing reading from e-read-          ables, recent work has discovered further factors that link to screen-
ers versus books (Chang, Aeschbach, Duffy, & Czeisler, 2015), this        time, such as self-regulation skills (Cliff et al., 2018) and factors lying
study adds mental imagery to this already complicated picture             behind familial and socioeconomic factors, such as stress. Although,
(Barr & Linebarger, 2017). Conceivably, media might be tailored           to our knowledge, work has not yet investigated whether these fac-
to also encourage mental imagery, or in educational settings to           tors link to mental imagery, the current findings need to be tempered
be embedded in other activities that stimulate the sensorimotor           by the fact that third variables may explain links found with screen-
system, such as activities that involve outdoor experiences.              time. Additionally, we utilized a single mental imagery measure
SUGGATE and MARTZOG                                                                                                                               |   11 of 13

including items targeting visual and haptic modalities. Subsequent             A systematic review. Developmental Review, 44, 19–58. https://doi.
                                                                               org/10.1016/j.dr.2016.12.002
work should examine a broader repertoire of mental imagery, includ-
                                                                           Anguera, J. A., Boccanfuso, J., Rintoul, J. L., Al-Hashimi, O., Faraji, F.,
ing imagery pertaining to other sensory modalities and motor imag-             Janowich, J., … Gazzaley, A. (2013). Video game training enhances
ery (Borst, 2014). Given links between screen-time and executive               cognitive control in older adults. Nature, 501(7465), 97–101. https://
functions (Lillard & Peterson, 2011), a more comprehensive battery             doi.org/10.1038/natur​e12486
                                                                           Arbuckle, J. L. (2014). Amos (Version 23.0) [Computer Program]. Chicago,
extending beyond working memory might serve as an additional
                                                                               IL: IBM SPSS.
control. Experimental work in which screen-time is manipulated,            Barr, R., & Linebarger, D. N. (2017). Media exposure during infancy and
controlling for media content and different levels of experience with          early childhood. Cham, Switzerland: Springer International Publishing.
screen media, could be used to compliment the more ecologically            Barsalou, L. W. (1999). Perceptions of perceptual symbols. Behavioral and
valid longitudinal cross-lagged panel design employed here.                    Brain Sciences, 22, 637–660. https://doi.org/10.1017/S0140​525X9​
                                                                               9532147
   Finally, our findings supported the reduction, but not the stim-
                                                                           Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology,
ulation, hypothesis. Concerning the latter, one might speculate that           59, 617–645. https://doi.org/10.1146/annur​                ev.psych.59.103006.
a general increase in processing speed in the visual system from in-           093639
creased screen-time, with its often rapidly changing visual images,        Beentjes, J. W. J., & van der Voort, T. H. A. (1988). Television's impact
                                                                               on children's reading skills: A review of research. Reading Research
might cause a general processing advantage (e.g., Dye et al., 2009),
                                                                               Quarterly, 23, 389–413. https://doi.org/10.2307/747640
perhaps leading also to a tendency to respond rapidly due to greater       Borst, G. (2014). Neural underpinnings of object mental imagery, spatial
impulsivity (e.g., Lillard & Peterson, 2011). However, our analyses            imagery, and motor imagery. In K. N. Ochsner & S. Kosslyn (Eds.), The
at best only resulted in a small, albeit expectedly negative, pathway          Oxford handbook of cognitive neuroscience, (Vol. 1, pp. 74–87). Core
                                                                               topics. Oxford, UK: Oxford University Press.
between screen-time and reaction time on the imagery task.
                                                                           Byrne, B. M. (2010). Structural equation modeling with AMOS: Basic con-
                                                                               cepts, applications, and programming (2nd ed.). Multivariate applica-
                                                                               tions series. New York, NY: Routledge.
5 | CO N C LU S I O N                                                      Chang, A.-M., Aeschbach, D., Duffy, J. F., & Czeisler, C. A. (2015). Evening
                                                                               use of light-emitting eReaders negatively affects sleep, circadian tim-
                                                                               ing, and next-morning alertness. Proceedings of the National Academy
The ability to draw forth mental images, inspecting and comparing              of Sciences of the United States of America, 112, 1232–1237. https://
these, would appear to be a foundational human faculty lying at                doi.org/10.1073/pnas.14184​90112
the heart of cognitive functioning (Kosslyn et al., 2003). On the          Chonchaiya, W., & Pruksananonda, C. (2008). Television viewing associ-
                                                                               ates with delayed language development. Acta Paediatrica, 97, 977–
one hand, society is becoming increasingly technical and special-
                                                                               982. https://doi.org/10.1111/j.1651-2227.2008.00831.x
ized, on the other hand, more dynamic and instable. Accordingly,           Cliff, D. P., Howard, S. J., Radesky, J. S., McNeill, J., & Vella, S. A. (2018).
education will need to ensure that children are creative and in-               Early childhood media exposure and self-regulation: Bidirectional
novative, alongside acquiring specialized skills. Mental imagery is            longitudinal associations. Academic Pediatrics, 18, 813–819. https://
precisely an ability that can contribute to novel problem solving              doi.org/10.1016/j.acap.2018.04.012
                                                                           Connell, L., & Lynott, D. (2012). Strength of perceptual experience pre-
and adaptive thinking as well as specialized skills, hence it would
                                                                               dicts word processing performance better than concreteness or im-
seem wise to avoid compromising children's development in this                 ageability. Cognition, 125, 452–465. https://doi.org/10.1016/j.cogni​
regard. In this sense, the current findings that screen-time nega-             tion.2012.07.010
tively affect mental imagery need to be actively replicated and            Connell, L., & Lynott, D. (2016). Embodied semantic effects in visual
                                                                               word recognition. In M. H. Fischer & Y. Coello (Eds.), Conceptual
pursued, with a focus on better understanding underlying mecha-
                                                                               and interactive embodiment (pp. 71–92). London, NewYork, NY:
nisms, potentially leading to interventions to reduce screen-media             Routledge.
usage (Schmidt et al., 2012), and participation in educational ac-         Dunckley, V. L. (2015). Screentime is making kids moody, crazy and lazy:
tivities to foster mental imagery.                                             6 ways electronic screen time makes kids angry, depressed and un-
                                                                               motivated. Retrieved fromhttps://www.psych​ology​today.com/intl/
                                                                               blog/menta ​ l -wealt ​ h /20150 ​ 8/scree ​ n time ​ - is-makin ​ g-kids-moody​
AC K N OW L E D G M E N T S                                                    -crazy​-and-lazy
This research was supported by a grant from the Software - AG              Dworak, M., Schierl, T., Bruns, T., & Strüder, H. K. (2007). Impact of
Stiftung awarded to the first author (grant number: ER-P 11657). We            singular excessive computer game and television exposure on
                                                                               sleep patterns and memory performance of school-aged children.
thank all participating children, parents, teachers, and schools and
                                                                               Pediatrics, 120, 978–985. https://doi.org/10.1542/peds.2007-
the research assistants who helped with data collection.                       0476
                                                                           Dye, M. W. G., Green, C. S., & Bavelier, D. (2009). Increasing speed of pro-
DATA AVA I L A B I L I T Y S TAT E M E N T                                     cessing with action video games. Current Directions in Psychological
                                                                               Science, 18, 321–326. https://doi.org/10.1111/j.1467-8721.2009.
The data that support the findings of this study are available from
                                                                               01660.x
the corresponding author upon reasonable request.                          Endlich, D., Berger, N., Küspert, P., Lenhard, W., Marx, P., Weber, J., &
                                                                               Schneider, W. (2017). Würzburger Vorschultest: Erfassung schriftspra-
REFERENCES                                                                     chlicher und mathematischer (Vorläufer-)Fertigkeiten und sprachlicher
Allchorne, K., Cooper, N. R., & Simpson, A. (2017). The relationship be-       Kompetenzen im letzten Kindergartenjahr [Würzburger preschool test].
    tween television exposure and children's cognition and behaviour:          Göttingen: Hogrefe.
12 of 13   |                                                                                                                         SUGGATE and MARTZOG

Ennemoser, M., & Schneider, W. (2007). Relations of television view-                    from the science of learning. Psychological Science in the Public
    ing and reading: Findings from a 4-year longitudinal study.                         Interest: A Journal of the American Psychological Society, 16, 3–34.
    Journal of Educational Psychology, 99, 349–368. https://doi.                        https://doi.org/10.1177/15291​0 0615​569721
    org/10.1037/0022-0663.99.2.349                                                 Hutton, J. S., Dudley, J., Horowitz-Kraus, T., DeWitt, T., & Holland,
Federal Bureau of Statistics [Statistisches Bundesamt]. (2017).                         S. K. (2019). Associations Between Screen-Based Media Use and
    Bildungsstand der Bevölkerung 2016 [Educationallevel of the pop-                    Brain White Matter Integrity in Preschool-Aged Children. JAMA
    ulation 2016] Retrieved from https://www.desta​             tis.de/DE/Publi​        Pediatrics, 174(1), e193869, https://doi.org/10.1001/jamap​         ediat​
    katio​nen/Thema​tisch​/Bildu​ngFor​schun​gKult​ur/Bildu​ngsst​and/Bildu​            rics.2019.3869
    ngsst​andBe​voelk​erung​52100​02167​0 04.pdf?blob=publi​c atio​nFile           James, K. H. (2010). Sensori-motor experience leads to changes in vi-
Feierabend, S., Plankenhorn, T., & Rathgeb, T. (2017). Kindheit, Internet               sual processing in the developing brain. Developmental Science, 13,
    Medien: Basisstudie zum Medienumgang 6–13-Jähriger in Deutschland.                  279–288. https://doi.org/10.1111/j.1467-7687.2009.00883.x
    Stuttgart, Germany: Medienpädagogischer Forschungsverbund                      Jeannerod, M. (2001). Neural simulation of action: A unifying mechanism
    Südwest (LFK, LMK).                                                                 for motor cognition. NeuroImage, 14(1 Pt 2), S103–S109. https://doi.
Fitzpatrick, C., Pagani, L. S., & Barnett, T. A. (2012). Early childhood televi-        org/10.1006/nimg.2001.0832
    sion viewing predicts explosive leg strength and waist circumference           Kaufman, A.S., & Kaufman, N.L. (2015). Kaufman assessment battery for
    by middle childhood. The International Journal of Behavioral Nutrition              children–second edition (KABC-2). Göttingen, Germany: Hogrefe.
    and Physical Activity, 9, 87. https://doi.org/10.1186/1479-5868-9-87           Kearney, M. W. (2017). Cross-Lagged Panel Analysis. In M. Allen (Ed.),
Galetzka, C. (2017). Commentary: Mobile and interactive media use by                    The SAGE encyclopedia of communication research methods. Thousand
    young children: The good, the bad, and the unknown. Frontiers in                    Oaks, CA, Boston, MA: SAGE Publications, Inc; Credo Reference.
    Psychology, 8, 461. https://doi.org/10.3389/fpsyg.2017.00461                   Kirkorian, H. L., Wartella, E. A., & Anderson, D. R. (2008). Media and
Gingold, J. A., Simon, A. E., & Schoendorf, K. C. (2014). Excess screen                 young children's learning. The Future of Children, 18, 39–61. https://
    time in US children: Association with family rules and alternative ac-              doi.org/10.1353/foc.0.0002
    tivities. Clinical Pediatrics, 53, 41–50. https://doi.org/10.1177/00099​       Kline, R. B. (2011). Principles and practice of structural equation modeling.
    22813​498152                                                                        Methodology in the social sciences (3rd ed.). New York, NY: Guilford
Glenberg, A., Brown, M., & Levin, J. (2007). Enhancing comprehension                    Press. Retrieved from http://site.ebrary.com/lib/acade​       micco​mplet​
    in small reading groups using a manipulation strategy. Contemporary                 etitl​es/home.action
    Educational Psychology, 32, 389–399. https://doi.org/10.1016/j.                Kosslyn, S. M. (1994). The resolution of the imagery debate. Cambridge,
    cedps​ych.2006.03.00                                                                UK: MIT Press.
Glenberg, A. M., & Gallese, V. (2012). Action-based language: A theory             Kosslyn, S. M., Ganis, G., & Thompson, W. L. (2001). Neural foundations
    of language acquisition, comprehension, and production. Cortex, 48,                 of imagery. Nature Reviews Neuroscience, 2, 635–642. https://doi.
    905–922. https://doi.org/10.1016/j.cortex.2011.04.010                               org/10.1038/35090055
Glenberg, A. M., Sato, M., Cattaneo, L., Riggio, L., Palumbo, D., & Buccino,       Kosslyn, S. M., Ganis, G., & Thompson, W. L. (2003). Mental imagery:
    G. (2008). Processing abstract language modulates motor system                      Against the nihilistic hypothesis. Trends in Cognitive Sciences, 7, 109–
    activity. Quarterly Journal of Experimental Psychology, 61, 905–919.                111. https://doi.org/10.1016/S1364​-6613(03)00025​-1
    https://doi.org/10.1080/17470​21070​1625550                                    Kosslyn, S. M., Margolis, J. A., Barrett, A. M., Goldknopf, E. J., & Daly, P.
Grush, R. (2004). The emulation theory of representation: Motor control,                F. (1990). Age differences in imagery abilities. Child Development, 61,
    imagery, and perception. Behavioral and Brain Sciences, 27, 377–396.                995. https://doi.org/10.2307/1130871
    https://doi.org/10.1017/S0140​525X0​4 000093                                   Krcmar, M., Grela, B., & Lin, K. (2007). Can toddlers learn vocabulary
Hancox, R. J., Milne, B. J., & Poulton, R. (2005). Association of televi-               from television? An experimental approach. Media Psychology, 10,
    sion viewing during childhood with poor educational achievement.                    41–63. https://doi.org/10.1080/15213​26070​1300931
    Archives of Pediatric Adolescent Medicine, 159, 614–618. https://doi.          Lewkowicz, D. J. (2000). The development of intersensory temporal
    org/10.1001/archp​edi.159.7.614                                                     perception: An epigenetic systems/limitations view. Psychological
Hargreaves, I. S., Pexman, P. M., Johnson, J. C., & Zdrazilova, L. (2012).              Bulletin, 126, 281–308. https://doi.org/10.1037//0033-2909.126.
    Richer concepts are better remembered: Number of features ef-                       2.281
    fects in free recall. Frontiers in Human Neuroscience, 6. https://doi.         Lillard, A. S., & Peterson, J. (2011). The immediate impact of different
    org/10.3389/fnhum.2012.00073                                                        types of television on young children's executive function. Pediatrics,
Hattie, J. (2012). Visible learning for teachers: Maximizing impact on learn-           128, 644–649. https://doi.org/10.1542/peds.2010-1919
    ing. London, UK; New York, NY: Routledge Taylor & Francis Group.               Linebarger, D. L., & Piotrowski, J. T. (2009). TV as storyteller: How ex-
Heitz, R. P. (2014). The speed-accuracy tradeoff: History, physiology,                  posure to television narratives impacts at-risk preschoolers' story
    methodology, and behavior. Frontiers in Neuroscience, 8, 150. https://              knowledge and narrative skills. British Journal of Developmental
    doi.org/10.3389/fnins.2014.00150                                                    Psychology, 27, 47–69. https://doi.org/10.1348/02615​1008x​4 00445
Herodotou, C. (2018). Young children and tablets: A systematic review              Linebarger, D. L., & Vaala, S. E. (2010). Screen media and language de-
    of effects on learning and development. Journal of Computer Assisted                velopment in infants and toddlers: An ecological perspective.
    Learning, 34, 1–9. https://doi.org/10.1111/jcal.12220                               Developmental Review, 30, 176–202. https://doi.org/10.1016/j.
Hinkley, T., Salmon, J., Okely, A. D., Crawford, D., & Hesketh, K. (2012).              dr.2010.03.006
    Preschoolers' physical activity, screen time, and compliance with              Linebarger, D. L., & Walker, D. (2005). Infants' and toddlers' television
    recommendations. Medicine and Science in Sports and Exercise, 44,                   viewing and language outcomes. American Behavioral Scientist, 48,
    458–465. https://doi.org/10.1249/MSS.0b013​e3182​33763b                             624–645. https://doi.org/10.1177/00027​6 4204​271505
Hipp, D., Gerhardstein, P., Zimmermann, L., Moser, A., Taylor, G., & Barr,         Madigan, S., Browne, D., Racine, N., Mori, C., & Tough, S. (2019).
    R. (2017). The dimensional divide: Learning from TV and touch-                      Association between screen time and children's performance on a
    screens during early childhood. In R. Barr & D. N. Linebarger (Eds.),               developmental screening test. JAMA Pediatrics, 173(3), 244. https://
    Media exposure during infancy and early childhood. Cham, Switzerland:               doi.org/10.1001/jamap​ediat​rics.2018.5056
    Springer International Publishing.                                             Martzog, P., & Suggate, S. (2019). Fine motor skills and mental imagery: Is
Hirsh-Pasek, K., Zosh, J. M., Golinkoff, R. M., Gray, J. H., Robb, M. B., &             it all in the mind? Journal of Experimental Child Psychology, 186, 59–72.
    Kaufman, J. (2015). Putting education in "educational" apps: Lessons                https://doi.org/10.1016/j.jecp.2019.05.002
You can also read